%20WHITE.png)
American Science Is About to Have Its Napster Moment. The Indie Music Scene Already Wrote the Survival Guide.
By Michael W. Nestor, Ph.D.
Michael W. Nestor, Ph.D., is a neuroscientist and longtime independent musician who founded a nonprofit musicians’ cooperative and indie label in Baltimore.
As AI, automation and concentrated funding reshape American research, neuroscientist and indie musician Michael W. Nestor argues that science is approaching the same rupture music faced after Napster—and that DIY culture offers a survival guide.
In 2002, I was doing two things that had no business happening at the same time. I was touring in an indie rock band and starting a neuroscience PhD. I spent my days in a lab and my nights loading amps into a van, watching the bottom fall out of the record business in real time.
A year later, I started a not-for-profit musicians’ cooperative and indie label in Baltimore because neither the major nor independent-label system was serving the people making the records.
I have spent the twenty years since inside science. Now I am watching some of the same pressures we dealt with in early-aughts rock reach research: concentration at the top, the disappearance of the working middle, an obsession with measurable output, and technologies that lower the cost of making things while making the people who make them easier to ignore.
The indie music world may also be a survival guide.
The middle is always the first to go
Everybody remembers Napster as the thing that killed the record industry. That is not quite what happened. The superstars were generally fine. The major labels shrank, merged, and survived. What largely disappeared was the middle, the working musician who could fill a 500-seat room, sell a few thousand records, and build a life without becoming famous.
People kept making music, but the support system around working bands was thinning out, record stores, distributors, local press, modest guarantees, and the record sales that helped pay for the next tour. By the time streaming became dominant, listeners had access to more music than ever while most musicians were earning less and less from their tours and their recordings.
What I recognize in the current American scientific moment is that same loss of a viable middle. In research, there are still stars with enormous labs and institutions built to keep winning, just as there are still musicians selling out arenas. But the path to becoming a working scientist with a lab of one’s own has narrowed. The average researcher now receives a first major independent grant at around 43, compared with about 35 in 1980, while roughly 10 percent of funded researchers receive 40 percent of the money.
Music still has its stars, and science still has its star labs. What both are losing is the working middle beneath them. It is the K-shaped economy manifest in both.
Loud, louder, loudest
You already know the loudness wars, even if you never called them that. Since the early 1990s, records kept getting mastered louder because louder tracks cut through more easily on radio and in record-store listening stations. The tradeoff was less dynamic range and flatter-sounding records.
My band did it too. In 2007, we released an album mastered at a higher loudness level than what we were comfortable with because that was the standard we were working inside. Later, we released a second version with the dynamic range restored to what we originally intended for the record. That record got caught up in and became part of a national conversation about the loudness wars. It was my introduction to the real influence the digital domain could have on the analog one outside of the standard arguments around the use of autotune or plugins for recording.
When streaming became the main way people heard music, the pressure moved beyond mastering. Artists had to think about plays, skips, playlist placement, song length, and release frequency. Those numbers began influencing the songs themselves, in how quickly the vocal entered, where the hook appeared, and how long the track ran.
These are versions of what economists call “Goodhart’s Law.” Once a measure becomes a target, people start changing the work to improve the number.
Science has built a similar system, much of it under the banner of metascience. In trying to measure scientific performance, the field has elevated citations, impact factors, publication totals, grant dollars, and institutional rankings. Those numbers were supposed to describe research. They have increasingly determined what gets funded, published, and rewarded.
Then came AI
AI and automation are usually discussed as a science and technology story. They are already a music story too.
In science, automated laboratories can run experiments around the clock, while machine-learning systems search huge datasets and propose compounds or hypotheses. A small group may soon be able to do work that once required a major institution.
Music has heard this promise before. Digital recording and file sharing lowered barriers, but a few centralized platforms emerged to control discovery. Almost anyone could release a song, leaving every musician to compete with millions of others.
Generative AI pushes this further by making nearly unlimited music available at almost no cost. Science faces the same risk: automated labs and AI-generated hypotheses may produce more research without making it easier to tell what matters.
AI can make songs but it cannot create the bands, venues, labels, writers, and listeners that turn those songs into a scene.
Automated labs can run experiments, but scientists still have to decide which questions are worth asking.
In both fields, production is becoming cheap. Judgment, creativity, and passion are priceless.
The money behind both systems
Underneath both stories is a problem of time.
Labels once expected some bands to take two or three records to break. As the business focused more on immediate returns, regional scenes, unusual records, and artists who grew slowly were the first to lose support.
Science is under similar pressure. Long-term research is difficult to finance because the payoff may be uncertain and many years away. Low interest rates made that easier after 2008, especially in biotechnology and other speculative fields. With capital now more expensive, investors are narrowing what they will fund and for how long.
Federal research policy is also moving toward milestones, portfolio reviews, shorter funding periods, and the ability to end projects when priorities change. Those tools may be useful for some applied research. The risk comes when they are used for work that cannot show value on that schedule.
Both systems narrow the window available to every artist or research program to show impact and force each into a game where each must prove itself early. Bands get less time to develop, and scientists move toward projects that can produce a result before the next review, increasing the temptation to cut corners or dismiss potentially transformative work as “too early.”
Music has already shown what follows: fewer strange bets, weaker scenes, and more work designed around what the system will reward quickly, with the audience losing out on real art.
What we DIY basement kids figured out early
What survived the collapse of the old record business? The bands and labels that had already learned to work outside it.
Labels ran out of apartments and group houses, mine included. Artists pressed their own records, booked tours, traded shows, held onto their masters, and built direct relationships with listeners. That DIY infrastructure still exists, although artists now carry much more of the cost themselves.
DIY meant taking on work the industry used to do. It was exhausting, but it also gave artists more control over the records, the audience, and the institutions they built around them.
Science now has its own versions: open-source tools, shared facilities, remote laboratories, preprints, independent research organizations, and direct public communication.
AI and automation could make those alternatives more viable. But music offers a warning. Musicians gained cheap recording and global distribution, then discovered that the platforms controlling attention had become more powerful than many of the labels they replaced.
Science could follow the same path. Researchers may gain powerful models and automated labs while becoming dependent on the companies that own the compute, datasets, and infrastructure. Better tools will not mean independence if researchers must give up control to use them.
Two scenes, one fight
From the inside, music and science look less different than they do from a distance.
Both spend years making things whose value may not be clear at first. Both rely on communities and institutions willing to support work before anyone knows whether it will succeed. And in both, new technology keeps getting sold as progress while it gets harder for the people doing the work to stay in the fight.
The indie scene learned that cheaper recording and easier distribution did not make the business fairer. It mostly moved the gatekeepers somewhere else.
I would like to see working musicians and scientists brought into the same room, not as a gimmick, but to compare what is happening to their work. Label founders who built direct relationships with listeners have something to teach researchers trying to work outside the grant system. Scientists building shared laboratories and open tools may have something to offer musicians confronting AI and another flood of nearly free content.
Science may be approaching the kind of break music went through at the turn of the century. AI, laboratory automation, concentrated funding, rising costs, and declining trust are not separate problems. Together, they are changing who can do the work and on what terms.
Napster did not end music. It made music easier to distribute while hollowing out the part of the business that allowed working bands to survive. We are still trying to rebuild that middle.
Science could end up with more papers, more data, and wider access while leaving fewer people able to build a career doing research. That is the Napster moment science should fear: a flood of research produced by a system increasingly unable to sustain the people doing it.
What happens to American science matters to music because both require a country willing to support work before its value is obvious. The country that stops making room for uncertain science will not keep making room for uncertain music. Anyone still trying to make a record, build a scene, or keep a band going should recognize the warning.
American science should too: once the middle is gone, you can keep counting outputs for a while, but the next generation of discovery has already been cut off.
Indie music has already shown American science what a country loses when the people in the middle can no longer afford to keep betting years of their lives on something that might matter.
